In view of the lack of traditional image segmentation algorithms to guide Convolutional Neural Network （CNN） for segmentation in the current field of medical image segmentation， a medical image segmentation Network with Content-Guided Multi-Angle Feature Fusion （CGMAFF-Net） was proposed. Firstly， grayscale images and Otsu threshold segmentation images were used to generate lesion region guidance maps through a Transformer-based micro U-shaped feature extraction module， and Adaptive Combination Weighting （ACW） was used to weight them to the original medical images for initial guidance. Then， Residual Network （ResNet） was employed to extract downsampled features from the weighted medical images， and a Multi-Angle Feature Fusion （MAFF） module was used to fuse feature maps at 1/16 and 1/8 scales. Finally， Reverse Attention （RA） was applied to upsample and restore the feature map size gradually， so as to predict key lesion regions. Experimental results on CVC-ClinicDB， Kvasir-SEG， and ISIC 2018 datasets demonstrate that compared to the existing best-performing segmentation multiscale spatial reverse attention network MSRAformer， CGMAFF-Net increases the mean Intersection over Union （mIoU） by 0.97， 0.78， and 0.11 percentage points， respectively； compared to the classic network U-Net， CGMAFF-Net improves the mIoU by 2.66， 8.94， and 1.69 percentage points， respectively， fully verifying the effectiveness and advancement of CGMAFF-Net.

Addressing the insufficiency of structural guidance in the framework of Zero-Shot Action Recognition （ZSAR） algorithms， an Action Recognition Algorithm based on Attention mechanism and Energy function （ARAAE） was proposed guided by the Energy-Based Model （EBM） for framework design. Firstly， to obtain the input for EBM， a combination of optical flow and Convolutional 3D （C3D） architecture was designed to extract visual features， achieving spatial non-redundancy. Secondly， Vision Transformer （ViT） was utilized for visual feature extraction to reduce temporal redundancy， and ViT cooperated with combination of optical flow and C3D architecture was used to reduce spatial redundancy， resulting in a non-redundant visual space. Finally， to measure the correlation between visual space and semantic space， an energy score evaluation mechanism was realized with the design of a joint loss function for optimization experiments. Experimental results on HMDB51 and UCF101 datasets using six classical ZSAR algorithms and algorithms in recent literature show that on the HMDB51 dataset with average grouping， the average recognition accuracy of ARAAE is （22.1±1.8）%， which is better than those of CAGE （Coupling Adversarial Graph Embedding）， Bi-dir GAN （Bi-directional Generative Adversarial Network） and ETSAN （Energy-based Temporal Summarized Attentive Network）. On UCF101 dataset with average grouping， the average recognition accuracy of ARAAE is （22.4±1.6）%， which is better than those of all comparison algorithm slightly. On UCF101 with 81/20 dataset segmentation method， the average recognition accuracy of ARAAE is （40.2±2.6）%， which is higher than those of the comparison algorithms. It can be seen that ARAAE improves the recognition performance in ZSAR effectively.

To solve the problems of image detail loss and unclear texture caused by interference factors such as noise， imaging technology and imaging principles in the medical Magnetic Resonance Imaging （MRI） process， a multi-receptive field generative adversarial network for medical MRI image super-resolution reconstruction was proposed. First， the multi-receptive field feature extraction block was used to obtain the global feature information of the image under different receptive fields. In order to avoid the loss of detailed texture due to too small or too large receptive fields， each set of features was divided into two groups， and one of which was used to feedback global feature information under different scales of receptive fields， and the other group was used to enrich the local detailed texture information of the next set of features； then， the multi-receptive field feature extraction block was used to construct feature fusion group， and spatial attention module was added to each feature fusion group to adequately obtain the spatial feature information of the image， reducing the loss of shallow and local features in the network， and achieving a more realistic degree in the details of the image. Secondly， the gradient map of the low-resolution image was converted into the gradient map of the high-resolution image to assist the reconstruction of the super-resolution image. Finally， the restored gradient map was integrated into the super-resolution branch to provide structural prior information for super-resolution reconstruction， which was helpful to generate high quality super-resolution images. The experimental results show that compared with the Structure-Preserving Super-Resolution with gradient guidance （SPSR） algorithm， the proposed algorithm improves the Peak Signal-to-Noise Ratio （PSNR） by 4.8%， 2.7% and 3.5% at ×2， ×3 and ×4 scales， respectively， and the reconstructed medical MRI images have richer texture details and more realistic visual effects.

Concerning the problem that uncertainty information is difficult to be merged during the decision-making process of multi-source irrigation information, a decision fusion method based on fuzzy rough set and Dempster-Shafer (D-S) evidence theory was proposed. Using the fuzzy rough set theory,the basic probability distribution function was established, the interdependence between irrigation factors and irrigation decision was calculated, and the identification framework of irrigation decision on the multiple fusion irrigation factors was built. Using the improved D-S evidence theory, the multi-source irrigation information was fused at the decision-making level, the expression and synthesis problems of uncertain information were solved. The information of winter wheat such as soil moisture, photosynthetic rate and stomatal conductance in north China was fused in irrigation decision by the application of the methods mentioned above. The results show that the uncertainty of the irrigation decision decreases from 38.0% before fusion to 9.84%. The method can effectively improve the accuracy of irrigation decision and reduce the uncertainty of the irrigation decision.

To address the too much time consumption issue in the selection of parameter values existing in the mixed kernel composed of traditional kernels, a method of constructing mixed kernel based on Fisher-kernel was proposed. Because of the non-parameter characteristic of Fisher-kernel, the number of parameters in the Fisher-kernel based mixed kernel was effectively reduced, thus the selection time of parameter values was also effectively reduced. The experimental results on typical color face databases show that, compared with traditional mixed kernel, the parameter selection time of Fisher-kernel based mixed kernel is significantly reduced and the correctness rate of recognition is improved, which confirms the effectiveness of the proposed method.

An echo cancellation algorithm based on additional signal was proposed to eliminate the spatial co-channel interference of digital repeater in mobile communication network. Firstly, the paper studied the formation and characteristics of the echo, and established the model of true echo channel parameter matrix. Secondly, the single frequency sinusoidal signal was attached to the frequency spectrum hole of the base station, and the channel parameters of the estimated-echo matrix were estimated by making use of the convolution relationship of the cross-correlation function of the mixed-signal received by repeater, the transponder signal and the additional signal. Finally, the estimated-echo of the transponder signal was subtracted from the received mixed signal, which was achieved by the estimated-echo channel, in order to eliminate the echo interference. The simulation results show that the relative error of the attenuation coefficient of main path is 1.8493×10-5 in the frequency selective fading channel model of COST 207 standard. The estimated-echo can track the actual-echo better, so as to avoid self-excited phenomenon effectively.

Concerning the shortcomings in reliability and limited bandwidth of conventional dual-threshold energy sensing for cognitive radio, this paper introduced a new dual-threshold collaborative spectrum sensing algorithm based on two-bit hard combination. The algorithm made use of two kinds of information which included one-bit local decisions and two-bit local decisions of secondary users for eliminating the sensing failure. Then fusion center made a final decision based on the two kinds of decisions to determine whether the primary user was present or not. Compared to the conventional dual-threshold algorithm, the simulation results indicate that the new algorithm not only eliminates the sensing failure, but also improves the detection performance significantly (maximum increasing by about 21% at low false alarm probabilities) at the cost of slightly increasing the average number of sensing bits (on average increasing by about 1% at low sensing failure probabilities).

To check the redundant services of Internet of Things (IoT) and save resources, a novel similarity calculation model based on service event class diagram was proposed to check redundancy by means of relations between events and services. It analyzed the context of IoT and service type and obtained the services similarity measurement based on events according to this model. By this similarity calculation method, a static service redundancy detection algorithm was proposed to remove duplication function invocation of services, which saved system resources occupied by services and decreased the consumption of resources in the IoT.

On the basis of fast Multi-Agent System (MAS) application development with Agent Pattern Language (APL), this paper proposed a framework of process-oriented enterprise application integration. According to the general characteristics of business process, the framework adopted a hierarchy structure for the business process to separate business logic and implementation. Then some local changes of sub-process had less influence on the execution of global business process. Besides, roles and structured activities were adopted in the framework. The business process can be divided into sub-process with roles to form the hierarchy structure. The structured activities were used to describe complex business process. With the framework, the enterprise application integration can be implemented in an effective way, and the implementation will be more flexible.

In order to make multiple disks accessible parallelly, the stripe layout is adopted by RAID for high-bandwidth. The approach is appropriate for big sequential data access, but is not that effective for small data random access. This paper analyzed the I/O performance based on RAID Stripe size, explored the IOPS（IOs per second）of multiple user small data access, and proposed a new stripe layout architecture—Thick Granularity Model. The simulation indicates: the I/O performance of Thick Granularity Model has been improved，especially in small data random access.

Considering the high cost of the traditional AODVjr algorithm in ZIGBEE MESH network, an improved algorithm was proposed which was based on the role differences and current energy state of the node. Some important or low-energy nodes can avoid being disabled for continuing excess energy consumption, which can cause route invalidation and even network disability. The simulation indicates that the algorithm can improve the transport reliability, reduce the energy consumption and prolong the lifetime of the network.

Query processing is one of the most important technologies in sensor networks. To reduce energy consumption of query, a small world based query strategy named Contact-Assisted poweR-efficient Direction-sense-achieved query strategy for Sensor Networks （CardSN） was presented. In CardSN, Contacts act as shortcuts to bring down the average path length of the network; a distributed relative localization algorithm was introduced to achieve a sense of direction. Experimental results show that CardSN has good scalability and can save more energy than ZRP and CAPTURE. Therefore, CardSN is of high-performance and energy-efficiency.

To meet the demand of high spatio-temporal resolution remote sensing images for power facilities safety monitoring and emergency management， a deep convolutional network-based remote sensing image fusion enhancement model for power facilities was proposed. Firstly， a deep convolutional network was designed， including encoder， Residual Attention （RA） mechanism block， substitution attention mechanism block and decoder. Secondly， the two-layer convolution and the residual block of fusion channel attention mechanism were improved to increase the network's attention to details and key features of images， and enhance the feature extraction capability of the network. Thirdly， the multi-channel substitution attention block was improved to make the network paying more attention to the details of images. As the result， the performance of high-resolution image fusion reconstruction was improved. Finally， the loss function composition of the model was improved， and the composite loss function consisting of content loss and visual loss was adopted to improve training effect of the model. Experimental results indicate that the proposed model has the performance of image fusion reconstruction better than other fusion models significantly， and the detail textures of predicted image closer to those of the real image. Compared with Multi-stage Feature Compensation NET （MFCNET） model， the proposed model has the Correlation Coefficient （CC） improved by 1.6%. and the SSIM （Structure Similarity Index Measure） improved by 18.4%. It can be seen that the proposed model provides a basis for remote sensing image processing， especially for high-resolution reconstruction of small target remote sensing images.